Sync: Fix renderer crash when network disconnects after upload

[gavande1]

Related issues

• Fixes STU-1180

Proposed Changes

• Added error handling in getPushProgressInfo to catch network errors gracefully
• Skip Sentry reporting for expected crossDomain network errors (offline scenarios)
• Prevent unhandled promise rejections when network goes offline

Testing Instructions

1. Start a new sync operation
2. Allow the upload step to complete successfully
3. Disconnect the network (e.g., turn off Wi‑Fi) during subsequent sync steps
4. Verify the renderer no longer crashes and handles the offline state gracefully
5. Verify that when the network reconnects, the sync operation can continue

Pre-merge Checklist

• Have you checked for TypeScript, React or other console errors?

[github-actions]

📊 Performance Test Results

Comparing a1359a0 vs trunk

site-editor

Metric	trunk	a1359a0	Diff	Change
load	8244.00 ms	7315.00 ms	-929.00 ms	🟢 -11.3%

site-startup

Metric	trunk	a1359a0	Diff	Change
siteCreation	9065.00 ms	9067.00 ms	+2.00 ms	🔴 0.0%
siteStartup	3959.00 ms	3954.00 ms	-5.00 ms	🟢 -0.1%

---

Results are median values from multiple test runs.

Legend: 🟢 Improvement (faster) | 🔴 Regression (slower) | ⚪ No change

[katinthehatsite]

Network errors (e.g., offline) should be handled gracefully

I am not sure we need to include this comment though. It seems clear from the code itself. What do you think?

[gavande1]

That is a good point. It makes sense to remove it, as the code itself is quite explanatory.

[gavande1]

Removed in a1359a0

[katinthehatsite]

What is the process of throwing the console.error in this case? What use would it have for the user?

[gavande1]

AFAIU, the import status is pooled at intervals, but it continues to do so even when the network goes offline. On the trunk, the renderer crashes due to the network being offline. This is the expected behavior; however, the API itself might return an error from the server. I thought it would be useful to log this for debugging purposes, especially to handle any unknown cases.

[gavande1]

Removed in a1359a0

[katinthehatsite]

I see, thanks for explaining!

I am wondering if we should just skip logging it to the console to not pollute the console with something that we know will cause the error e.g. network interruption. What do you think?

Not a strong preference, just a suggestion on my end.

[gavande1]

Thanks for the suggestion Kate. I have removed logging after you suggestion.

[katinthehatsite]

It seems it works as described but when I was testing this, the sync progress did not indicate to me in any way that I was offline. Perhaps we could consider improving this area to be more indicative that the user is disconnected:

What do you think?

[katinthehatsite]

It seems that the GH approved the changes but I was not intending to approve them just yet, sorry about that

[gavande1]

It seems it works as described but when I was testing this, the sync progress did not indicate to me in any way that I was offline. Perhaps we could consider improving this area to be more indicative that the user is disconnected:

Excellent observation. I added the offline indicator in #2212, but it only appears during the upload phase and when the network goes offline. I will create a follow-up issue to address this.

[gavande1]

@katinthehatsite I have created STU-1195 linear issue to improve offline statuses during different phases of push.

[katinthehatsite]

@katinthehatsite I have created STU-1195 linear issue to improve offline statuses during different phases of push.

Nice sounds good to me 👍

[gavande1]

It seems that the GH approved the changes but I was not intending to approve them just yet, sorry about that

@katinthehatsite Do you mind giving another review?

[katinthehatsite]

Looks good on my end 👍 Thanks for implementing the adjustments!